CN111437519A - Multi-line beam selection method with optimal biological effect - Google Patents

Multi-line beam selection method with optimal biological effect Download PDF

Info

Publication number
CN111437519A
CN111437519A CN202010259006.6A CN202010259006A CN111437519A CN 111437519 A CN111437519 A CN 111437519A CN 202010259006 A CN202010259006 A CN 202010259006A CN 111437519 A CN111437519 A CN 111437519A
Authority
CN
China
Prior art keywords
tumor
layer unit
hidden layer
error
optimal
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202010259006.6A
Other languages
Chinese (zh)
Other versions
CN111437519B (en
Inventor
袁双虎
李玮
李莉
韩毅
刘宁
马志祥
袁朔
吕慧颖
于金明
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Jinan Bishan Network Technology Co ltd
Shandong Cancer Hospital & Institute (shandong Cancer Hospital)
Shandong University
Original Assignee
Jinan Bishan Network Technology Co ltd
Shandong Cancer Hospital & Institute (shandong Cancer Hospital)
Beijing Yikang Medical Technology Co ltd
Shandong University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Jinan Bishan Network Technology Co ltd, Shandong Cancer Hospital & Institute (shandong Cancer Hospital), Beijing Yikang Medical Technology Co ltd, Shandong University filed Critical Jinan Bishan Network Technology Co ltd
Priority to CN202010259006.6A priority Critical patent/CN111437519B/en
Publication of CN111437519A publication Critical patent/CN111437519A/en
Application granted granted Critical
Publication of CN111437519B publication Critical patent/CN111437519B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61N—ELECTROTHERAPY; MAGNETOTHERAPY; RADIATION THERAPY; ULTRASOUND THERAPY
    • A61N5/00—Radiation therapy
    • A61N5/10—X-ray therapy; Gamma-ray therapy; Particle-irradiation therapy
    • A61N5/103—Treatment planning systems
    • A61N5/1031—Treatment planning systems using a specific method of dose optimization

Abstract

The invention discloses a method for selecting a multi-line beam with optimal biological effect. The invention utilizes different characteristics of the multi-energy beam, monitors the change of the position and the shape of the tumor in the radiotherapy process in real time, statistically analyzes and searches the optimal ray and the dose distribution mode with the optimal biological effect, ensures the treatment state with the optimal biological effect in the whole radiotherapy process, and improves the tumor control rate.

Description

Multi-line beam selection method with optimal biological effect
Technical Field
The invention belongs to a multi-line beam selection method for radiotherapy with optimal biological effect, and particularly relates to the field of radiation physics.
Background
Radiotherapy is one of the main local treatment methods for malignant tumors, about 70% of cancer patients need radiotherapy, and 40% of cancer patients can achieve radical treatment effect through radiotherapy. The radiation sources commonly used in radiotherapy include radiation generated by radioisotopes and high-energy x-rays, electron beams, proton beams, neutron beams and other particle beams generated by various accelerators. The radiation therapy usually used includes both long-distance and short-distance radiation therapy, the radiation must pass through normal tissue to reach the tumor tissue during the long-distance radiation therapy, the dose of the tumor radiation is limited by the dose of the normal tissue around the tumor radiation, and the radiation with different energy and the multi-field radiation technology are required.
Tumors grow without a fixed shape, and malignant tumors are generally poorly differentiated and irregularly shaped. The high-energy X-ray has unique physical characteristics, the maximum dose is clearly positioned at a certain depth under the skin, after the area is built, the dose exponentially decays along with the increase of the tissue depth, and the percentage depth dose increases along with the increase of the ray dose, the irradiation field area and the source skin distance, so that the uniform distribution of the high-dose area of the whole tumor in the three-dimensional direction cannot be ensured. The morphology of individual tumors also changes as radiotherapy progresses, requiring real-time adjustment of the target volume, revised dose, or irradiation pattern, since small morphological changes may move high dose areas into surrounding normal tissue that should not be irradiated, and precise targeting may become precisely damaged or missed.
Because of this, the radiation therapy does not have the best biological effect throughout the treatment, and therefore a multi-beam selection of the best biological effect is performed.
Disclosure of Invention
In order to overcome the above-mentioned deficiencies of the prior art, the present invention provides a multi-beam selection method with optimal biological effect, which utilizes the advantages of multi-beam irradiation to make and adjust the multi-beam radiation dose and plan in real time by monitoring the tumor morphology changes during the radiation therapy in real time, so that the tumor tissue has the optimal dosimetry distribution during the whole radiation therapy.
The purpose of the invention can be realized by the following technical scheme:
a method for selecting a multi-line beam with optimal biological effect, comprising:
firstly, a Cone beam CT (Cone Beam computer tomogry, CBCT) machine is used for real-time online acquisition and monitoring of the position and the shape of a tumor focus in the radiotherapy process;
secondly, calculating the distance from a radiation source to the center of the tumor tissue in different three-dimensional states in real time according to the tumor shape monitored by the CBCT system, and irradiating the tumor tissue by radiation beams of different energy grades (6,8,10 and 12MeV) according to the depth table of the maximum dose point and the 50 percent dose point of the high-energy X-ray;
high energy X-ray maximum dose point and 50% dose point depth (cm)
Figure BDA0002438569790000021
Thirdly, converting the acquired data of the distance from the radioactive source to the center of the focus and the change of the radioactive ray beam into corresponding dose distribution data, and drawing a distance-dose distribution curve chart and a fitting function relation on a computer to obtain an optimal ray selection and dose distribution mode.
Fourthly, inputting the summarized optimal ray with the optimal biological effect and the dose distribution function into the radiotherapy system, applying the optimal ray and the dose distribution function to the next radiotherapy treatment course, and timely adjusting the scheme according to the treatment effect.
The method specifically comprises the following steps:
firstly, determining correct positioning according to surface marks, tumor images and the distribution of surrounding healthy visceral organs, performing multi-angle scanning irradiation by adopting a CBCT (cone beam computed tomography) machine, and acquiring and monitoring the position and the shape of a tumor focus on line in real time;
secondly, extracting the characteristics of the tumor focus region by using an algorithm based on wavelet transform, and classifying the characteristic values of the tumor image extracted by using the method by using a neural network to obtain distance change data from a radioactive source to a tumor tissue center;
thirdly, according to the distance change data acquired in the earlier stage, the radiation beams corresponding to different energy levels (6,8,10 and 12MeV) are converted into corresponding radiation dose distribution data; analyzing according to 70% of the obtained limited ray dose distribution data points to obtain a function with optimal ray selection and dose distribution, and performing function verification by using the remaining 30% of the data points to ensure that the error is small enough and the function has the optimal biological effect of the multi-energy ray;
and fourthly, inputting the function determined by verification into an operating system for clinical application.
Compared with the prior art, the invention has the beneficial effects that:
the invention carries out multi-energy beam scanning in real time according to the change of the position and the shape of the tumor focus in the radiotherapy process, analyzes and calculates the optimal ray and the dose distribution mode with the optimal biological effect, and is beneficial to improving the success rate of radiotherapy and the recovery rate of patients.
Drawings
FIG. 1 is a flow chart of a radiation therapy method of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all embodiments, and all other embodiments obtained by a person of ordinary skill in the art without creative efforts based on the embodiments of the present invention belong to the protection scope of the present invention.
Referring to fig. 1, the present invention provides a technical solution: a multi-line bundle selection method with optimal biological effect concretely comprises the following steps:
(1) the CBCT machine is adopted for multi-angle scanning irradiation, the wavelet transform-based algorithm is used for extracting the characteristics of the tumor focus region, and the specific process is as follows:
by using
Figure BDA0002438569790000031
Representing a tumor image expansion function
Figure BDA0002438569790000032
And scale factor s:
Figure BDA0002438569790000041
function f (x) continuous wavelet transform at scale s and x:
Figure BDA0002438569790000042
to ensure the reconstruction characteristics of the tumor image data, the scale parameters are sampled into a binary sequence {2 }j}j∈ZTo obtain a sequence of dyadic wavelet transform functions
Figure BDA0002438569790000043
For image data normalization, assume an optimal scale of 1 and a maximum scale of 2jSince the scale of the tumor image is limited, there are:
Figure BDA0002438569790000044
defining a discrete sequence of preferential energies as fdThen, the following relationship is obtained:
Figure BDA0002438569790000045
thus, any dimension 2 can be obtainedJDiscrete signal sequence
Figure BDA0002438569790000046
The characteristic parameters of the tumor image are obtained through the transformation.
(2) The tumor image characteristic values extracted by the method are classified by using a neural network, the extracted characteristic values are learned through a neuron model, and then the extracted characteristic values are classified. The neural network comprises an input layer, a hidden layer and an output layer, neurons in all layers are connected with one another completely, and neurons in all layers are not connected with one another, and the specific classification learning algorithm is as follows:
connection right V for input layer unit to hidden layer unithi(h 1, 2.. times, n), (i 1, 2.. times, p), the hidden layer unit to output layer unit connection weight Wij(i 1, 2.. times.p), (j 1, 2.. times.q) and a threshold θ of the hidden layer unitiThreshold assignment of element of output layerjAssigning a random value in the interval (-1, + 1);
for sample mode (A)k，Ck) (k ═ 1, 2.., m) the following operations were performed:
① the value of A is sent to the input layer unit and the hidden layer unit via the connection weight matrix V to generate the activation value of the hidden layer unit
Figure BDA0002438569790000051
Where i is 1, 2, and p, f is a sigmoid function:
Figure BDA0002438569790000052
② calculating activation values for output layer elements
Figure BDA0002438569790000053
③ calculating the generalized error of the output layer unit
Figure BDA0002438569790000054
Wherein j is 1, 2., q,
Figure BDA0002438569790000055
is the desired output of output layer unit j.
④ calculating hidden layer elements for each djError of (2)
Figure BDA0002438569790000056
Wherein i is 1, 2,. and p; the above equation is equivalent to back-propagating the error of the output layer unit to the hidden layer;
⑤ adjusting the connection weight of hidden layer to output layer
Δwij＝λbidj
Wherein i is 1, 2, p, i is 1, 2, q, λ is learning rate, 0 < λ < 1.
⑥ adjusting the connection weight of input layer to hidden layer
Δvhi＝βakei
Wherein h is 1, 2, 1, n and i is 1, 2, p, 0 < β < 1.
⑦ adjusting threshold of output layer unit
Δrj＝λdj
Wherein j is 1, 2
⑧ adjusting the threshold of the hidden layer cell
Δθj＝βei
Wherein i is 1, 2
Repeating step (2) until an error d of m for k 1, 2jBecomes sufficiently small or becomes 0. The algorithm described above performs the mapping of the input to the output by a process that minimizes the error function.
(3) According to the characteristics, the distance from the radiation source to the center of the tumor is obtained in a multi-angle three-dimensional state, and the radiation beams corresponding to different energy levels (6,8,10,12MeV) are converted into corresponding radiation dose distribution data; analysis was performed based on 70% of the limited ray dose distribution data points obtained, resulting in a function with the best ray selection and dose distribution, and function validation was performed using the remaining 30% data points.
(4) And inputting the determined function into an operating system according to verification for clinical application to obtain the multi-line beam selection method with the optimal biological effect.
Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.

Claims (3)

1. A method for multi-line beam selection with optimal biological effect, comprising the steps of:
1) determining correct positioning according to the surface mark, the tumor image and the distribution of the surrounding healthy visceral organs, performing multi-angle scanning irradiation by adopting a CBCT (cone beam computed tomography) machine, and acquiring and monitoring the position and the shape of a tumor focus on line in real time;
2) extracting the characteristics of a tumor focus region by using an algorithm based on wavelet transform, and classifying the tumor image characteristic values extracted by using the method by using a neural network to obtain distance change data from a radioactive source to a tumor tissue center;
3) according to the distance change data acquired in the earlier stage, the radiation beams corresponding to different energy levels are converted into corresponding radiation dose distribution data; and analyzing 70% of the limited ray dose distribution data points to obtain a function with the optimal ray selection and dose distribution, and performing function verification by using the remaining 30% of the data points to ensure that the error is small enough and the optimal biological effect of the multi-energy ray is achieved.
2. The method according to claim 1, wherein the wavelet transform-based algorithm extracts the features of the tumor lesion region by the following steps:
by using
Figure FDA0002438569780000011
Representing a tumor image expansion function
Figure FDA0002438569780000012
And scale factor s:
Figure FDA0002438569780000013
function f (x) continuous wavelet transform at scale s and x:
Figure FDA0002438569780000014
to ensure the reconstruction characteristics of the tumor image data, the scale parameters are sampled into a binary sequence {2 }j}j∈ZTo obtain a sequence of dyadic wavelet transform functions
Figure FDA0002438569780000015
For image data normalization, assume an optimal scale of 1 and a maximum scale of 2jSince the scale of the tumor image is limited, there are:
Figure FDA0002438569780000021
defining a discrete sequence of preferential energies as fdThen, the following relationship is obtained:
Figure FDA0002438569780000022
thus, any dimension 2 can be obtainedJDiscrete signal sequence
Figure FDA0002438569780000023
The characteristic parameters of the tumor image are obtained through the transformation.
3. The method of claim 1, wherein the extracted tumor image feature values are classified by a neural network, the extracted feature values are learned by a neuron model, and then the extracted feature values are classified, the neural network comprises an input layer, a hidden layer and an output layer, the neurons in each layer are fully connected, the neurons in each layer are not connected, and the specific classification learning algorithm is as follows:
connection right V for input layer unit to hidden layer unithi(h 1, 2.. times, n), (i 1, 2.. times, p), the hidden layer unit to output layer unit connection weight Wij(i 1, 2.. times.p), (j 1, 2.. times.q) and a threshold θ of the hidden layer unitiThreshold assignment of element of output layerjAssigning a random value in the interval (-1, + 1);
for sample mode (A)k，Ck) (k ═ 1, 2.., m) the following operations were performed:
① the value of A is sent to the input layer unit and the hidden layer unit via the connection weight matrix V to generate the activation value of the hidden layer unit
Figure FDA0002438569780000024
Where i is 1, 2, and p, f is a sigmoid function:
Figure FDA0002438569780000025
② calculating activation values for output layer elements
Figure FDA0002438569780000031
③ calculating the generalized error of the output layer unit
Figure FDA0002438569780000032
Wherein i is 1, 2., q,
Figure FDA0002438569780000033
is the desired output of output layer unit j;
④ calculating hidden layer elements for each djError of (2)
Figure FDA0002438569780000034
Wherein i is 1, 2,. and p; the above equation is equivalent to back-propagating the error of the output layer unit to the hidden layer;
⑤ adjusting the connection weight of hidden layer to output layer
Δwij＝λbidj
Wherein i is 1, 2, 1, p, j is 1, 2, 1, q, λ is learning rate, 0 < λ < 1;
⑥ adjusting the connection weight of input layer to hidden layer
Δvhi＝βakei
Wherein h is 1, 2, 1, n and i is 1, 2, p, 0 < β < 1;
⑦ adjusting threshold of output layer unit
Δrj＝λdj
Wherein j is 1, 2
⑧ adjusting the threshold of the hidden layer cell
Δθj＝βei
Wherein i is 1, 2
Repeating step (2) until an error d of m for k 1, 2jUntil it becomes sufficiently small or 0, the algorithm performs input to output mapping by a process that minimizes the error function.
CN202010259006.6A 2020-04-03 2020-04-03 Multi-line beam selection method with optimal biological effect Active CN111437519B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010259006.6A CN111437519B (en) 2020-04-03 2020-04-03 Multi-line beam selection method with optimal biological effect

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010259006.6A CN111437519B (en) 2020-04-03 2020-04-03 Multi-line beam selection method with optimal biological effect

Publications (2)

Publication Number Publication Date
CN111437519A true CN111437519A (en) 2020-07-24
CN111437519B CN111437519B (en) 2021-10-19

Family

ID=71652796

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010259006.6A Active CN111437519B (en) 2020-04-03 2020-04-03 Multi-line beam selection method with optimal biological effect

Country Status (1)

Country Link
CN (1) CN111437519B (en)

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8467874B2 (en) * 2000-04-11 2013-06-18 The Board Of Regents Of The University Of Texas System Gastrointestinal electrical stimulation
CN107403438A (en) * 2017-08-07 2017-11-28 河海大学常州校区 Improve the ultrasonoscopy focal zone dividing method of fuzzy clustering algorithm
CN110559009A (en) * 2019-09-04 2019-12-13 中山大学 Method, system and medium for converting multi-modal low-dose CT into high-dose CT based on GAN

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8467874B2 (en) * 2000-04-11 2013-06-18 The Board Of Regents Of The University Of Texas System Gastrointestinal electrical stimulation
CN107403438A (en) * 2017-08-07 2017-11-28 河海大学常州校区 Improve the ultrasonoscopy focal zone dividing method of fuzzy clustering algorithm
CN110559009A (en) * 2019-09-04 2019-12-13 中山大学 Method, system and medium for converting multi-modal low-dose CT into high-dose CT based on GAN

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
翁邓胡: "基于CT图像的三维电子线放射治疗计划系统的研究与实现", 《中国优秀硕士学位论文全文数据库（电子期刊）》 *
郭凌云: "基于小波变换的乳腺肿瘤超声图像识别研究", 《中国优秀硕士学位论文全文数据库（电子期刊）》 *

Also Published As

Publication number Publication date
CN111437519B (en) 2021-10-19

Similar Documents

Publication Publication Date Title
EP3357537B1 (en) Geometric model establishment method based on medical image data
EP3369458B1 (en) Method for evaluating irradiation angle of beam
CN101120871A (en) Precise radiotherapy planning system
WO2021237869A1 (en) Parameter monitoring device and system for proton therapy
US11087524B2 (en) Method for establishing smooth geometric model based on data of medical image
CN108778420A (en) Neutron capture cure treatment planning systems
CN103394167A (en) BED (biological effective dose)-based prediction method for complications caused by tumor radiotherapy
CN107666940A (en) The method for selecting beam geometry
CN111951926B (en) Method for detecting sensitivity of different types of tumors to radiotherapy rays based on machine learning technology
CN107292075A (en) Promote the method that radiotherapy system calculates benefit
US20230111230A1 (en) Radiotherapy system and treatment plan generation method therefor
CN111437519B (en) Multi-line beam selection method with optimal biological effect
EP4166193A1 (en) Radiotherapy system and therapy plan generation method therefor
CN112263788A (en) Quantitative detection system for morphological change in radiotherapy process
CN105105780A (en) Method and device for generating control point of dynamic wedge-shaped plate
CN115445102A (en) Boron neutron capture treatment method and device
CN113426030B (en) Proton dosage calculation method and device
Huntelerslag Modelling Dosimetric Variations Due to Beam Delivery Parameter Perturbations
Sandnes Comparison of relative biological effectiveness in passive scattering-and pencil beam scanning proton therapy of pediatric cancer
Unnikrishnan et al. Monte carlo simulation of electron beams from varian truebeam linear accelerator used in radiotherapy: estimation of initial beam parameters
CN111790065A (en) Multi-beam combined radiotherapy method for tumor
CN117727423A (en) Particle beam range real-time verification method and system
Høeg ROBUSTNESS ANALYSIS OF RADIATION TREATMENT PLANS FOR BREAST CANCER, A COMPARATIVE STUDY BETWEEN PROTONS AND PHOTONS
CN114822766A (en) Rapid photon proton dose prediction and screening algorithm
CN116312954A (en) Intensity modulated radiotherapy plan virtual verification method

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
TA01 Transfer of patent application right
TA01 Transfer of patent application right

Effective date of registration: 20210924

Address after: No. 440, Yankuang Road, Ji'nan, Shandong Province

Applicant after: SHANDONG CANCER HOSPITAL & INSTITUTE (SHANDONG CANCER Hospital)

Applicant after: SHANDONG University

Applicant after: Jinan Bishan Network Technology Co.,Ltd.

Address before: 100025 906, West District, 9th floor, No. 100, Sili, Balizhuang, Chaoyang District, Beijing

Applicant before: Beijing Yikang Medical Technology Co.,Ltd.

Applicant before: SHANDONG CANCER HOSPITAL & INSTITUTE (SHANDONG CANCER Hospital)

Applicant before: SHANDONG University

Applicant before: Jinan Bishan Network Technology Co.,Ltd.

GR01 Patent grant
GR01 Patent grant